Abstract
Background
Rare diseases such as pulmonary arterial hypertension (PAH) require centralized referral centres for diagnosis and management. It is unclear whether a patient’s geographic proximity to a referral centre influences the severity of illness on diagnosis and follow-up.
Methods
We identified all patients with a diagnosis of PAH across 3 tertiary referral centres. Postal code data were used to stratify patients based on community size. Commute time to a tertiary centre was determined using mapping software. Community size and commute time were compared with baseline catheterization data in addition to follow-up data including biomarkers, functional class, 6-minute walk distance, echocardiography, and repeat catheterization, if available. Follow-up data were used to create a risk factor stratification score as per the most recent guidelines of the European Society of Cardiology/European Respiratory Society.
Results
Of the 504 patients in our analysis, 51 (10.1%) lived in a rural area, 69 (13.7%) in a small population centre, 60 (11.9%) in a medium population centre, and 324 (64.3%) in a large urban centre. For commute time, 308 (61.1%) patients had a commute time < 1 hour, 93 (18.5%) patients commuted 1-3 hours, and 103 (20.4%) of patients commuted > 3 hours. There were no robust or consistent associations found between community size or commute time and baseline right heart catheterization data or follow-up indices.
Conclusions
In this multicentre analysis we found no clear associations between community size or commute time with severity of illness at diagnosis or follow-up. This supports previous conclusions that patients living in rural communities do not experience disparities in their PAH outcomes.
Keywords: community size, health policy, outcomes, population, pulmonary arterial hypertension
Résumé
Contexte
Le diagnostic et la prise en charge des maladies rares comme l’hypertension artérielle pulmonaire (HTAP) relèvent d’établissements spécialisés centralisés. On ignore si la proximité géographique d’un tel établissement influe sur la gravité de la maladie au moment du diagnostic et du suivi.
Méthodologie
Nous avons recensé tous les patients ayant reçu un diagnostic d’HTAP dans trois centres tertiaires spécialisés et utilisé le code postal pour les stratifier en fonction de la taille de l’agglomération. Nous avons ensuite déterminé le temps nécessaire pour se rendre dans un centre tertiaire à l’aide d’un logiciel de cartographie. La taille de l’agglomération et la durée du trajet ont été comparées aux données de référence se rapportant au cathétérisme, ainsi qu’aux données de suivi se rapportant aux biomarqueurs, à la classe fonctionnelle, au test de marche de 6 minutes, à l’échocardiographie et, le cas échéant, à la réalisation d’un second cathétérisme. Les données de suivi ont servi à créer un score de stratification du facteur de risque, conformément aux lignes directrices de l’ERS/ESC (European Respiratory Society/European Society of Cardiology) les plus récentes.
Résultats
Sur les 504 patients de notre analyse, 51 (10,1 %) résidaient en zone rurale, 69 (13,7 %) dans une petite agglomération, 60 (11,9 %) dans une agglomération de taille moyenne et 324 (64,3 %) dans un grand centre urbain. La durée du trajet était inférieure à 1 heure pour 308 patients (61,1 %), variait de 1 à 3 heures pour 93 patients (18,5 %) et était supérieure à 3 heures pour 103 patients (20,4 %). Il n’y avait aucune corrélation claire ou constante entre la taille de l’agglomération ou la durée du trajet et les données de référence relatives au cathétérisme du cœur droit ou aux paramètres de suivi.
Conclusions
Cette analyse multicentrique n’a mis en évidence aucune corrélation claire entre la taille de l’agglomération ou la durée du trajet et la gravité de la maladie au moment du diagnostic ou du suivi. Ces résultats confirment les conclusions des travaux antérieurs, à savoir qu’il n’y a pas de disparités dans les mesures de l’HTAP entre les patients vivant en zone rurale et les patients des zones urbaines.
Pulmonary arterial hypertension (PAH) is a progressive disease resulting from endothelial dysfunction and vascular remodeling of the small pulmonary arteries, which results in increased pulmonary vascular resistance (PVR), compromised right heart function, and death if untreated.1 Diagnosis requires right heart catheterization (RHC), which can be misleading if not done by experienced operators. Patient outcomes are improved when diagnosis is established and treatment provided at specialized expert centres, and this is the standard of care.2, 3, 4 The importance of earlier diagnosis and management in improving outcomes for these patients was emphasized in the most recent guidelines of the European Society of Cardiology/European Respiratory Society (ESC/ERS).5 Of concern is whether concentrating expertise in specialized centres delays diagnosis and treatment for those patients who live in rural or remote environments, as late referral to expert centres remains common.3 This is of particular importance in a country with a large land mass like Canada, where rural population centres are often remotely located from expert centres.
The effect of community size and commute time on severity of illness at diagnosis and at follow-up was previously explored in a group of 342 patients diagnosed with World Health Organization (WHO) Group 1 (PAH) or Group 4 (due to pulmonary artery obstructions) pulmonary hypertension (PH) across British Columbia, Canada.6 In this analysis, no associations were found between community size or commute time with severity of illness at diagnosis or prognostic markers on follow-up. Although these findings were reassuring, the results were limited by the small sample size and the single-centre nature of the analysis. Furthermore, when considering other illnesses, previous research has demonstrated discrepancies in outcomes and/or access to care based on geographic location in left-sided heart failure,7 respiratory illness including chronic obstructive pulmonary disease (COPD) and lung cancer,8 as well as cardiovascular disease.9 Given those findings, in addition to the expected heterogeneity in care delivery and population distribution in the different jurisdictions, multicentre analyses are clearly needed.
Our goal in this study was to determine whether community size or commute time was associated with severity of illness at diagnosis or on follow-up in patients with WHO Group 1 PAH across multiple diverse Canadian referral centres.
Methods
Data used in this this study were obtained from the following 3 main referral centres across Canada via the Canadian Pulmonary Hypertension Registry (CPHR): Vancouver, British Columbia; Calgary, Alberta; and Hamilton, Ontario. All patients met diagnostic criteria for PAH on RHC. Diagnostic RHCs are performed in designated tertiary centres by a preferred operator to maintain consistency. Furthermore, in Canada, PAH medication prescription is restricted to expert centres. This study was approved by the research ethics board of the University of British Columbia (H19-00657). Patients enrolled up until 2021 were included in analysis. We included patients diagnosed with PAH at the 3 clinic sites fulfilling hemodynamic criteria at the time of study proposal (mean pulmonary artery pressure [mPAP] > 20 mm Hg, wedge pressure < 15 mm Hg, PVR > 3 WU).10 Among the patients reviewed, PAH was predominantly diagnosed from 2010 to 2020.
Canadian census data were used to identify community size based on postal code data.11 Communities were classified as rural (population < 1000), small (population 1000-29,999), medium (population 30,000-99,999), or large (population > 100,000). Postal codes were also used to calculate the estimated commute time from patient residence to respective PH clinic using a web mapping service (Google Maps). The shortest commute time by car was used. If patients resided across bodies of water, commute time by ferry was factored into the travel time.
We collected initial demographic data for each patient, including date of birth, postal code, community size, body mass index (BMI), and sex. Key initial RHC data at diagnosis were also collected, including mPAP, cardiac index (CI), PVR, mean right atrial pressure (mRAP), and mixed venous oxygen saturation (SVO2). Mortality data were not available from a large enough subset of patients to be included in the analysis.
Follow-up data used in the analysis included 6-minute walk distance, WHO functional class, brain natriuretic peptide (BNP)/N-terminal proBNP, follow-up RHC data (if available), and pharmacotherapy (combination and parenteral therapy regimens, with combination referring to 2 or more oral medications, or oral and parenteral combinations). Select echocardiographic parameters were also evaluated, such as tricuspid annular plane systolic excursion (TAPSE), right atrial (RA) area, presence of pericardial effusion, and right ventricular systolic pressure. Data for risk stratification were collected from the last available complete follow-up. We required follow-up data to be at least 3 months from the initial diagnostic RHC. Patients were monitored at both via in person and telehealth appointments, as appropriate.
In our analysis we looked at severity of PAH on RHC at diagnosis and assessed risk status using the ESC/ERS guidelines at follow-up. Patients were assigned low- (0), intermediate- (1), or high-risk (2) status, according to the approach described by Kylhammar and colleagues.12 Risk score data were classified as a rounded average, as well as the percentage of patients achieving low-risk status. The variables included were WHO functional class, 6-minute walk distance, BNP/NT-proBNP, and follow-up echocardiogram and catheterization parameters. In accordance with 2022 guidelines, we used the updated cutoff scores for BNP/NT-proBNP as well as the addition of TAPSE/systolic pulmonary artery pressure (sPAP).5,10 Only current follow-up echocardiogram and catheterization data were included in the risk score. Our primary outcome was the proportion of patients achieving low risk. Missing data were simply omitted from the analysis. The risk score was calculated only with the variables available at the clinic appointment, as is commonly done in clinical practice.
Continuous variables are presented as mean ± standard deviation and categorical variables are presented as frequency and percentage. The Kruskal-Wallis test was used for all comparisons of continuous variables across the 3 clinics (Calgary, Hamilton, and Vancouver), including baseline hemodynamic measurements (mRAP, sPAP, mPAP, CI, PVR, and SVO2) and follow-up measurements. This nonparametric approach was chosen to maintain consistency and because it is robust to deviations from normality. For categorical variables, we used the χ2 test of independence to compare distributions across the 3 clinics.
To investigate the association between community size or commute time (both categorized) and each of the parameters at baseline or follow-up, we conducted multivariate linear or logistic regression analyses depending on the nature of the outcome variable. Our primary approach used mixed-effects models to account for the potential clustering effect of the 3 clinics. As a sensitivity analysis, we also performed generalized estimating equations and naive normal regression models. To account for differences in patient characteristics across sites, clinic was included as a random effect and patients were clustered by site to control for clinic-level variability.
The multivariate analyses adjusted for age, sex, and BMI as potential confounders. This approach was necessary due to sample size limitations and to avoid overadjustment, which could lead to reduced statistical power and increased standard errors.
This statistical approach provided a robust method for comparing the 3 clinics and investigating the impact of community size and commute time on patient outcomes, while accounting for potential confounding factors and the clustered nature of the data.
All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
Results
We identified 504 participants who met the inclusion criteria for this study. Baseline demographics and cardiac catheterization data for all participants, stratified by clinic site, are shown in Table 1. The median (q1-q3) time between initial RHC and follow-up echocardiogram was 1016 (468-1809) days. The median (q1-q3) time between initial RHC and follow-up RHC was 610 (217-1220) days. Four patients (1%), all from the Hamilton clinic, were missing initial RHC data. Of the remaining patients from whom initial RHC data were available, the vast majority (93%-99%) had data for mRAP, sPAP, mPAP, CI, and PVR. However, SVO2 was only collected in 47% of participants from the initial RHC. In terms of follow-up data, New York Heart Association (NYHA) data were only missing from 29 patients (5.8%), 6MWD data were only missing from 107 patients (21.2%), and most patients (75.2%) had one of either BNP or NT-proBNP. Far fewer patients had follow-up echocardiographic or follow-up RHC data because those investigations were not routinely ordered for all patients. However, because we used a rounded average of the available data from each participant, a risk score was obtained for most (95.0%) patients in the analysis. Baseline catheterization data as a function of community size and commute time are shown in Table 2. On average, patients from Hamilton demonstrated more favourable hemodynamics at baseline, with lower mRAP, mPAP, and PVR and higher CI when compared with other sites. Figure 1, A-C shows the distribution of patients across British Columbia, Alberta, and Ontario, respectively. The figures further demonstrate the number of patients residing in each city or town as well as the driving distance to their respective PH centre.
Table 1.
Baseline demographics and cardiac catheterization data for all participants, stratified by clinic
| Clinic |
Total (N = 504) | P value | |||
|---|---|---|---|---|---|
| Calgary (n = 111) | Hamilton (n = 133) | Vancouver (n = 260) | |||
| Age | 59.9 (15.7) | 60.7 (16.7) | 55.1 (15.6) | 57.6 (16.1) | 0.0006 |
| Female, n (%) | 71 (64.0%) | 97 (72.9%) | 198 (76.2%) | 366 (72.6%) | 0.0544 |
| BMI (kg/m2) | 29.5 (8.0) | 29.9 (7.8) | 27.6 (6.7) | 28.6 (7.4) | 0.0076 |
| Community size, n (%) | 0.1781 | ||||
| Rural | 14 (12.6%) | 9 (6.8%) | 28 (10.8%) | 51 (10.1%) | |
| Small | 13 (11.7%) | 16 (12.0%) | 40 (15.4%) | 69 (13.7%) | |
| Medium | 16 (14.4%) | 10 (7.5%) | 34 (13.1%) | 60 (11.9%) | |
| Large urban | 68 (61.3%) | 98 (73.7%) | 158 (60.8%) | 324 (64.3%) | |
| Commute time, n (%) | < 0.0001 | ||||
| < 1 hour | 84 (75.7%) | 109 (82.0%) | 115 (44.2%) | 308 (61.1%) | |
| 1-3 hours | 20 (18.0%) | 24 (18.0%) | 49 (18.8%) | 93 (18.5%) | |
| > 3 hours | 7 (6.3%) | 0 (0.0%) | 96 (36.9%) | 103 (20.4%) | |
| mRAP (mmHg) | 9.2 (5.5) | 7.2 (4.8) | 9.0 (5.0) | 8.6 (5.2) | 0.0020 |
| sPAP (mmHg) | 74.0 (21.1) | 72.4 (18.9) | 73.9 (20.0) | 73.5 (20.0) | 0.5043 |
| mPAP (mmHg) | 46.6 (12.7) | 43.5 (12.6) | 46.3 (12.2) | 45.7 (12.5) | 0.0258 |
| CI (L/min/m2) | 2.3 (0.7) | 2.6 (0.8) | 2.4 (0.7) | 2.4 (0.7) | 0.0122 |
| PVR (WU) | 9.3 (4.7) | 8.0 (5.2) | 9.1 (5.1) | 8.9 (5.1) | 0.0158 |
| SVO2(%) | 63.0 (9.0) | 63.3 (9.6) | 63.5 (10.1) | 63.2 (9.5) | 0.8929 |
Notes: Data are expressed as mean (standard deviation) unless otherwise specified. Continuous variables were compared using the Kruskal-Wallis test and categorical variables were compared using the χ2 test.
BMI, body mass index; CI, cardiac index; mPAP, mean pulmonary artery pressure; mRAP, mean right atrial pressure; PVR, pulmonary vascular resistance; sPAP, systolic pulmonary artery pressure; SVO2, mixed venous oxygen saturation.
Table 2.
Baseline cardiac catheterization data for all participants, stratified by commute time and community size
| Commute time |
|||||
|---|---|---|---|---|---|
| a) < 1 hour (n = 308) |
b) 1-3 hours (n = 93) |
c) > 3 hours (n = 103) |
Total (n = 504) | P value | |
| mRAP (mmHg) | 8.6 (5.2) | 8.7 (5.1) | 8.6 (4.9) | 8.6 (5.2) | 0.8946 |
| sPAP (mmHg) | 72.5 (20.4) | 75.1 (18.7) | 75.2 (19.8) | 73.5 (20.0) | 0.3040 |
| mPAP (mmHg) | 44.9 (13.1) | 46.0 (11.2) | 47.6 (11.3) | 45.7 (12.5) | 0.0600 |
| CI (L/min/m2) | 2.4 (0.7) | 2.4 (0.7) | 2.4 (0.8) | 2.4 (0.7) | 0.8931 |
| PVR (WU) | 8.6 (5.0) | 9.4 (5.5) | 9.2 (4.8) | 8.9 (5.1) | 0.3077 |
| SVO2(%) | 63.0 (9.3) | 63.0 (10.2) | 65.6 (9.5) | 63.2 (9.5) | 0.7381 |
| Community size |
||||||
|---|---|---|---|---|---|---|
| 1) Rural (n = 51) |
2) Small (n = 69) |
3) Medium (n = 60) |
4) Large urban (n = 324) |
Total (N = 504) | P value | |
| mRAP (mmHg) | 8.5 (4.9) | 9.6 (4.7) | 8.8 (5.6) | 8.4 (5.2) | 8.6 (5.2) | 0.1545 |
| sPAP (mmHg) | 75.6 (16.8) | 77.1 (21.9) | 71.0 (23.6) | 72.9 (19.2) | 73.5 (20.0) | 0.4634 |
| mPAP (mmHg) | 46.6 (10.1) | 47.7 (13.4) | 45.7 (13.5) | 45.0 (12.4) | 45.7 (12.5) | 0.3618 |
| CI (L/min/m2) | 2.3 (0.6) | 2.4 (0.7) | 2.5 (0.8) | 2.4 (0.8) | 2.4 (0.7) | 0.8672 |
| PVR (WU) | 9.7 (4.9) | 9.0 (5.2) | 8.6 (5.0) | 8.8 (5.1) | 8.9 (5.1) | 0.4521 |
| SVO2(%) | 63.5 (9.9) | 62.9 (8.5) | 63.8 (11.3) | 63.2 (9.3) | 63.2 (9.5) | 0.9664 |
Notes: Data are expressed as mean (standard deviation) unless otherwise specified. Continuous variables were compared using the Kruskal-Wallis test.
CI, cardiac index; mPAP, mean pulmonary artery pressure; mRAP, mean right atrial pressure; PVR, pulmonary vascular resistance; sPAP, systolic pulmonary artery pressure; SVO2, mixed venous oxygen saturation.
Figure 1.
Place of residence in relation to the pulmonary hypertension clinic in (A) Vancouver, British Columbia; (B) Edmonton, Alberta; and (C) Hamilton, Ontario.
There was no significant association between community size or commute time and baseline catheterization variables for any of the sites. When comparing follow-up clinical, echocardiographic, and RHC data stratified by community size and commute time, including respective risk scores, no significant differences were found apart from RA area by community size (P = 0.03). However, despite a significant P value in this case, the trend itself did not illustrate a consistent directional change in RA area with increasing community size (Table 3). Furthermore, using a more stringent significance threshold of P < 0.01 to account for multiple comparisons, this result was not considered statistically significant. In terms of our primary outcome variable of risk status, no robust differences were seen between the 3 sites at follow-up. There was also no significant relationship between ERS/ESC risk score and community size or commute time.
Table 3.
Follow-up clinical, echocardiogram, cardiac catheterization, and risk scores for all participants, stratified by commute time and community size
| Commute time |
Total (N = 504) | P value | |||
|---|---|---|---|---|---|
| a) < 1 hour (n = 308) |
b) 1-3 hours (n = 93) |
c) > 3 hours (n = 103) |
|||
| Combination therapy, n (%) | 167 (54.2%) | 56 (60.2%) | 59 (57.3%) | 282 (56.0%) | 0.5672 |
| Parenteral therapy, n (%) | 29 (9.4%) | 10 (10.8%) | 12 (11.7%) | 51 (10.1%) | 0.7889 |
| NYHA III-IV, n (%) | 137 (47.6%) | 45 (51.1%) | 39 (39.4%) | 221 (46.5%) | 0.2343 |
| 6MWD (m) | 362.1 (142.3) | 357.8 (137.3) | 401.9 (117.0) | 369.6 (137.1) | 0.0444 |
| TAPSE (mm) | 19.2 (6.1) | 19.4 (6.6) | 20.9 (4.0) | 19.6 (5.9) | 0.2679 |
| RA area (cm2) | 21.3 (8.4) | 21.0 (6.4) | 20.8 (5.6) | 21.1 (7.4) | 0.9222 |
| Pericardial effusion, n (%) | 0.8076 | ||||
| None | 173 (76.2%) | 53 (80.3%) | 56 (77.8%) | 282 (77.3%) | |
| Mild/minimum | 46 (20.3%) | 12 (18.2%) | 15 (20.8%) | 73 (20.0%) | |
| Moderate/severe | 8 (3.5%) | 1 (1.5%) | 1 (1.4%) | 10 (2.7%) | |
| RVSP (mmHg) | 60.3 (24.1) | 60.8 (24.4) | 60.9 (22.3) | 60.5 (23.7) | 0.9299 |
| BNP (pg/mL) | 275.7 (521.0) | 382.3 (824.8) | 186.5 (260.8) | 274.1 (554.7) | 0.6207 |
| NT-proBNP (pg/mL) | 3024.6 (6206.7) | 2456.6 (6006.1) | 1693.8 (2630.3) | 2801.3 (5939.7) | 0.6722 |
| RAP (mmHg) | 7.4 (4.8) | 8.1 (5.0) | 8.6 (6.5) | 7.8 (5.3) | 0.6682 |
| CI (L/min/m2) | 2.9 (0.8) | 3.5 (1.3) | 3.0 (0.8) | 3.1 (0.9) | 0.2448 |
| SVO2(%) | 64.3 (7.8) | 63.1 (9.0) | 65.8 (7.2) | 64.5 (7.8) | 0.5578 |
| Low risk, n (%) | 107 (36.9%) | 28 (31.5%) | 45 (45.0%) | 180 (37.6%) | 0.1477 |
| Community size |
Total (N = 504) | P value | ||||
|---|---|---|---|---|---|---|
| 1) Rural (n = 51) |
2) Small (n = 69) |
3) Medium (n = 60) |
4) Large urban (n = 324) |
|||
| Combination therapy, n (%) | 32 (62.7%) | 38 (55.1%) | 29 (48.3%) | 183 (56.5%) | 282 (56.0%) | 0.4887 |
| Parenteral therapy, n (%) | 5 (9.8%) | 7 (10.1%) | 3 (5.0%) | 36 (11.1%) | 51 (10.1%) | 0.5550 |
| NYHA III-IV, n (%) | 20 (40.0%) | 28 (43.1%) | 31 (52.5%) | 142 (47.2%) | 221 (46.5%) | 0.5567 |
| 6MWD (m) | 385.5 (140.5) | 371.8 (140.8) | 369.9 (133.0) | 366.2 (137.0) | 369.6 (137.1) | 0.8906 |
| TAPSE (mm) | 21.3 (5.3) | 18.7 (5.1) | 19.5 (5.7) | 19.5 (6.2) | 19.6 (5.9) | 0.1608 |
| RA area (cm2) | 21.7 (6.6) | 20.4 (6.2) | 18.0 (5.2) | 21.7 (8.0) | 21.1 (7.4) | 0.0304 |
| Pericardial effusion, n (%) | 0.7407 | |||||
| None | 32 (80.0%) | 39 (79.6%) | 32 (80.0%) | 179 (75.8%) | 282 (77.3%) | |
| Mild/minimum | 8 (20.0%) | 9 (18.4%) | 8 (20.0%) | 48 (20.3%) | 73 (20.0%) | |
| Moderate/severe | 0 (0.0%) | 1 (2.0%) | 0 (0.0%) | 9 (3.8%) | 10 (2.7%) | |
| RVSP (mmHg) | 62.2 (24.3) | 59.8 (23.6) | 60.9 (23.8) | 60.4 (23.8) | 60.5 (23.7) | 0.9693 |
| BNP (pg/mL) | 227.2 (306.5) | 345.5 (926.4) | 349.9 (585.6) | 253.4 (467.0) | 274.1 (554.7) | 0.6764 |
| NT-proBNP (pg/mL) | 2042.7 (4822.4) | 3220.7 (6428.9) | 1285.3 (3705.9) | 3076.5 (6270.8) | 2801.3 (5939.7) | 0.0493 |
| RAP (mmHg) | 8.9 (6.7) | 8.2 (6.2) | 6.3 (2.8) | 7.8 (5.2) | 7.8 (5.3) | 0.8305 |
| CI (L/min/mm2) | 2.8 (0.8) | 2.8 (0.8) | 3.3 (1.0) | 3.1 (1.0) | 3.1 (0.9) | 0.5386 |
| SVO2(%) | 63.8 (6.4) | 65.3 (9.7) | 66.1 (9.8) | 64.3 (7.5) | 64.5 (7.8) | 0.5053 |
| Low risk, n (%) | 19 (38.0%) | 25 (37.3%) | 27 (45.8%) | 109 (36.0%) | 180 (37.6%) | 0.5676 |
Notes: Data are expressed as mean (standard deviation) unless otherwise specified. Continuous variables were compared using the Kruskal-Wallis test and categorical variables were compared using the χ2 test.
6MWD, 6-minute walk distance; BNP, brain natriuretic peptide; CI, cardiac index; NT-proBNP, N-terminal pro–brain natriuretic peptide; NYHA, New York Heart Association; RAP, right atrial pressure; RVSP, right ventricular systolic pressure; TAPSE, tricuspid annular plane systolic excursion; SVO2, mixed venous oxygen saturation.
Using mixed-effects model analysis, no strong or consistent associations were identified between community size or commute time and baseline RHC data, as well as follow-up clinical, echocardiographic, or hemodynamic outcomes. Generalized estimating equations and naive normal regression models were also used as a sensitivity analysis, which did not demonstrate any meaningful relationships, and are not presented here for the sake of brevity. Given that the same sample was used to explore multiple outcome variables, the likelihood of identifying statistically significant results by chance alone is increased. To account for this, a more stringent threshold for significance (eg, P < 0.001) was applied. Under this criterion, no meaningful associations were observed. These results are included in Supplemental Tables S1-S4. None of the models indicated a statistically significant random intercept for clinic, suggesting that mean levels of the outcomes did not vary significantly between clinics. Some commute times were particularly long for the Vancouver clinic, reflecting its status as the sole provincial clinic in British Columbia. A simple linear regression model in the subset of patients with commute times > 3 hours showed no significant linear relationship between commute time and baseline mPAP and follow-up risk score.
Discussion
In our multicentre analysis of PAH referral centres operating in a public health care system with a geographically dispersed population, there were no robust or consistent relationships between community size or commute time and severity of disease on baseline catheterization data. Moreover, there was no relationship between these parameters and markers of successful treatment at follow-up, including achievement of low-risk status. These results persisted after adjusting for age, sex, BMI, and clinic effect.
Although research in other conditions, including left-sided heart failure, COPD, lung cancer, and cardiovascular disease, has demonstrated discrepancies between urban and rural communities,7, 8, 9 recent research in PH populations has demonstrated mixed results, and was not carried out in public health care systems. Our research expands on a previous single-centre cohort study,6 which did not demonstrate any association between community size or commute time with severity of illness at diagnosis or with markers of prognosis at follow-up. Our findings are congruent with those from a 2011 abstract examining 638 patients in the United States diagnosed with PAH, 234 of whom resided > 50 miles from a PH centre. Those patients showed no significant difference in WHO functional class, likelihood of delayed diagnosis, or time to initiation of prostacyclin analogs when compared with patients residing < 50 miles from a PH centre.13 In contrast, a recent study from the United States showed that the PH-related age-adjusted mortality rate between 2004 and 2019 was higher in rural vs urban counties.14 However, the analysis did not control for possible confounders, including baseline demographics, risk factors, and medical comorbidities. Furthermore, the only outcome variable was mortality rate, while additional data, such as echocardiography, RHC, biomarkers, and functional class, were not studied. Although the liberal use of telehealth could be proposed as a mechanism for the reassuring outcomes noted in our results, recent analyses have shown that rural populations may be less likely to use telemedicine services.15,16
It is important to acknowledge the presence of baseline demographic differences between sites, including a statistically significant difference in age, BMI, and commute time. This is reflective of the expected baseline differences in the populations that each expert centre serves. These baseline differences did not translate into differences stratified by community size or commute time. Therefore, it is likely that these differences in the descriptive data were largely accounted for by fixed effects (age, BMI). In addition, when the regression model was run while including clinic as a clustering effect, although trends were identified in certain isolated follow-up variables (CI and RA area), the overall lack of significance over all strata for community size and commute time suggests that these findings may be due to chance alone.
Another consideration of this study is the fact that hemodynamic criteria at the time of study proposal (mPAP > 20 mm Hg, wedge pressure <15 mm Hg, PVR > 3 WU),10 have since been updated to a lower PVR threshold of 2 WU.17 The data included in our study predate this revision, and patients with very mild PAH still make up a very small proportion of our registry. In the future it will be important to use revised definitions to ensure patients with subtle findings early in their disease course are being referred appropriately in both rural and urban environments.
There are some limitations to this study. First, although our sample size is large by PAH standards, patients living in very remote areas were underrepresented in some of the included sites. Second, the number of comorbidities included as covariates in the analysis was modest. It is possible that other factors (smoking history, hypertension, congestive heart failure, obstructive sleep apnea, COPD) may have influenced the results. However, to the point just noted, in the context of an already small sample size, introducing multiple covariates increases the risk of model complexity, multicollinearity, and an unreliable estimate. Furthermore, only a select number of echocardiographic variables were available in the CPHR for analysis. Additional markers of RV function, including strain imaging, would have been informative, but were not available. Finally, the retrospective nature of our analysis is an additional limitation from a data quality perspective, with the inability to establish any causality.
The strengths of our study include use of a relatively large sample size in the context of a rare disease such as PAH. Furthermore, broadening our previous analysis to include additional clinic sites served multiple purposes, including improving external validity, ensuring representation of a more diverse population served by 3 very different health boards and catchment area sizes, and enhancing insights to a more heterogeneous population.
There are many ways to build on this research. Ongoing expansion to additional sites, in addition to increasing the number of patients in our current cohort over time, would provide a larger sample size and may elucidate trends that are currently not visible. In addition, moving toward the 4-strata model when analyzing follow-up data may also serve a similar purpose. Although it would have been beneficial to include chronic thromboembolic pulmonary hypertension (Group 4) patients, which were included in work by Brunner and colleagues,6 data from this patient cohort were only available from 2 of the 3 clinics. Along with the addition of more clinic sites, future research would ideally include patients with chronic thromboembolic pulmonary hypertension, as their clinical trajectory is often similar to those with PAH. It would also be of benefit to broaden the current analysis to acknowledge the social determinants of health, which include factors such as socioeconomic status, race, ethnicity, social support, culture, and language, access to care, and residential environment,18 as it is understood that these variables are closely linked to overall health and mortality risk. From the perspective of socioeconomic status, a recent analysis performed on a PAH cohort in Canada demonstrated that for every $1000 increase in after tax income per single person equivalent, there was a trend toward lower mPAP and PVR in addition to a significantly higher likelihood of low-risk classification.19 It may be of benefit to incorporate data from the Canadian Index of Multiple Deprivation (CIMD), which is an area-based, geographically derived index of deprivation and marginalization, into future analyses.20 Finally, even contemporary studies have shown that the time from symptom onset to diagnosis of PAH remains approximately 2 years.21 Further evaluations of techniques for optimizing PAH diagnosis in general, such as using artificial intelligence, novel diagnostics, or systems interventions, are clearly needed.
Conclusions
In our multicentre analysis of 3 major tertiary referral centres, there were no consistent or robust relationships between commute time or community size and severity of PAH at diagnosis based on RHC data or severity of disease on follow-up. Although reassuring, expanding on the current research to include additional sites across North America and internationally would provide further insight into possible discrepancies in care for patients living in rural communities.
Ethics Statement
The research reported in this article was approved by the UBC Research Ethics Board (REB, H19-00657).
Patient Consent
The authors confirm that patient consent is not applicable to this article, as this is a retrospective case report using de-identified data.
Funding Sources
This project received funding from Janssen, United Therapeutics, Bayer, Merck Sharp & Dohme, and Canada’s Drug Agency (formerly the Canadian Agency for Drugs and Technologies in Health).
Disclosures
The authors have no conflicts of interest to disclose.
Footnotes
See page 1058 for disclosure information.
To access the supplementary material accompanying this article, visit CJC Open at https://www.cjcopen.ca/ and at https://doi.org/10.1016/j.cjco.2026.04.014.
Supplementary Material
References
- 1.Lai Y.C., Potoka K.C., Champion H.C., Mora A.L., Gladwin M.T. Pulmonary arterial hypertension: the clinical syndrome. Circ Res. 2014;115:115–130. doi: 10.1161/CIRCRESAHA.115.301146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sahay S., Melendres-Groves L., Pawar L., Cajigas H.R. Pulmonary hypertension care center network: improving care and outcomes in pulmonary hypertension. Chest. 2017;151:749–754. doi: 10.1016/j.chest.2016.10.043. [DOI] [PubMed] [Google Scholar]
- 3.Deaño R.C., Glassner-Kolmin C., Rubenfire M., et al. Referral of patients with pulmonary hypertension diagnoses to tertiary pulmonary hypertension centers: the multicenter RePHerral study. JAMA Intern Med. 2013;173:887–893. doi: 10.1001/jamainternmed.2013.319. [DOI] [PubMed] [Google Scholar]
- 4.Saunders H., Helgeson S.A., Abdelrahim A., et al. Comparing diagnosis and treatment of pulmonary hypertension patients at a pulmonary hypertension center versus community centers. Diseases. 2022;10:5. doi: 10.3390/diseases10010005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Humbert M., Kovacs G., Hoeper M.M., et al. 2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Heart J. 2022;43:3618–3731. doi: 10.1093/eurheartj/ehac237. [DOI] [PubMed] [Google Scholar]
- 6.Brunner N.W., Legkaia L., Al-Ahmadi F., et al. Does community size or commute time affect severity of illness at diagnosis or quality of care in a centralized care model of pulmonary hypertension? Int J Cardiol. 2021;332:175–181. doi: 10.1016/j.ijcard.2021.03.035. [DOI] [PubMed] [Google Scholar]
- 7.Gamble J.M., Eurich D.T., Ezekowitz J.A., et al. Patterns of care and outcomes differ for urban versus rural patients with newly diagnosed heart failure, even in a universal healthcare system. Circ Heart Fail. 2011;4:317–323. doi: 10.1161/CIRCHEARTFAILURE.110.959262. [DOI] [PubMed] [Google Scholar]
- 8.Goodridge D., Lawson J., Duggleby W., et al. Health care utilization of patients with chronic obstructive pulmonary disease and lung cancer in the last 12 months of life. Respir Med. 2008;102:885–891. doi: 10.1016/j.rmed.2008.01.007. [DOI] [PubMed] [Google Scholar]
- 9.Chen E.W., Varghese M.S., Wu W.C. Geographic access to cardiac rehabilitation and cardiovascular disease outcomes: a study of rural United States counties. Curr Probl Cardiol. 2024;49 doi: 10.1016/j.cpcardiol.2024.102740. [DOI] [PubMed] [Google Scholar]
- 10.Simonneau G., Montani D., Celermajer D.S., et al. Haemodynamic definitions and updated clinical classification of pulmonary hypertension. Eur Respir J. 2019;53 doi: 10.1183/13993003.01913-2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Statistics Canada. Dictionary, census of population 2021. 2021.
- 12.Kylhammar D., Kjellström B., Hjalmarsson C., et al. A comprehensive risk stratification at early follow-up determines prognosis in pulmonary arterial hypertension. Eur Heart J. 2018;39:4175–4181. doi: 10.1093/eurheartj/ehx257. [DOI] [PubMed] [Google Scholar]
- 13.Markin C.J., Roessel L.L., Lai G.P., Turner M., Barst R.J. 18 Does geographic distance from a pulmonary hypertension center delay diagnosis and treatment? A REVEAL registry analysis. J Heart Lung Transplant. 2011;30(suppl):S14. [Google Scholar]
- 14.Issa R., Minhas A.M.K., Issa R., et al. Urban-rural disparities in pulmonary hypertension-related mortality between 2004 and 2019: a call to improve access to specialty care centers for rural residents in the United States. Curr Probl Cardiol. 2023;48 doi: 10.1016/j.cpcardiol.2023.101623. [DOI] [PubMed] [Google Scholar]
- 15.DeBolt C.L., Rheuban K.S., Harris D. Telehealth services: improving equity in pulmonary health outcomes. Clin Chest Med. 2023;44:651–660. doi: 10.1016/j.ccm.2023.03.017. [DOI] [PubMed] [Google Scholar]
- 16.Lee M., Tsai M.H., Tillewein H., Luckey G.S. Rural-urban disparities in telemedicine use among U.S. adults with cancer. Telemed e-Health. 2024;30:1539–1548. doi: 10.1089/tmj.2023.0572. [DOI] [PubMed] [Google Scholar]
- 17.Kularatne M., Gerges C., Jevnikar M., Humbert M., Montani D. Updated clinical classification and hemodynamic definitions of pulmonary hypertension and its clinical implications. J Cardiovasc Dev Dis. 2024;11:78. doi: 10.3390/jcdd11030078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Havranek E.P., Mujahid M.S., Barr D.A., et al. Social determinants of risk and outcomes for cardiovascular disease: a scientific statement from the American Heart Association. Circulation. 2015;132:873–898. doi: 10.1161/CIR.0000000000000228. [DOI] [PubMed] [Google Scholar]
- 19.Alquraishi H., Swiston J., Lee L., et al. D21 risk factors and predictors in critical illness and chronic respiratory diseases. The association between median income and severity of pulmonary hypertension at diagnosis and risk at follow up in a public health care system. Poster discussion session. ATS. 2022 https://www.atsjournals.org Available from: [Google Scholar]
- 20.Canadian Index of Multiple Deprivation: User guide. 2021. https://www150.statcan.gc.ca/n1/pub/45-20-0001/452000012023002-eng.htm Available from:
- 21.DuBrock H.M., Germack H.D., Gauthier-Loiselle M., et al. Economic burden of delayed diagnosis in patients with pulmonary arterial hypertension (PAH) Pharmacoecon Open. 2024;8:133–146. doi: 10.1007/s41669-023-00453-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.